Friendly Locals and Clean Streets?—Evaluating Jakarta’s Destination Brand Image

Round 1

Reviewer 1 Report

Interesting article, very carefully written and methodologically refined. Carefully carried out statistical identification. The value of the study is increased by the application nature of the study and conclusions that can be used by public administration and institutions of the tourist service sector to implement specific projects aimed at improving the image of the city among visitors.

Author Response

Thank you very much for the comments that have been given by the first reviewer. The comprehensive and meaningful comments are very much appreciated.

Reviewer 2 Report

Thank you for the opportunity to review the paper titled Friendly Locals and Clean Streets? – Evaluating Jakarta’s Destination Brand Image.

Overall, I find the paper clearly structured and coherent. The topic is relatively “classic” in the sense that there had indeed been written a lot on destination image and is, therefore, hard to provide new and original work, although the topic is highly important for each of the destinations for practical reasons. This is thus the prime weakness of this paper: the lack of originality.

This lack, however, might be compensated by adding two important points that would improve the paper’s solidness in the sense for it to be an example of “typical” research that would have practical benefits for destination marketers.

The first is that the literature review is relatively dated – I propose to update with newer references and to add a section on what all of the results mean now from the perspective of the COID-19. Due to this, I suspect that the survey was made in the pre-COVID times - I propose that the authors add the timeframe of collecting the data and discuss this.

The second is that the hypotheses and the model are presented like this is an explanatory model. In reality, the hypotheses are really about the relation of “whole – and part” – I propose that the authors make this point stronger and perhaps make this clearer also from the visualizations.

Author Response

Thank you for reviewing our paper titled Friendly Locals and Clean Streets? Evaluating Jakarta’s Destination Brand Image. Your valuable inputs are very much appreciated. Several revisions have been made according to the suggestions.

The first suggestion is to incorporate the current COVID-19 pandemic into the paper. We have added discussion of the results from the perspective of COVID-19 on page 10. Furthermore, we have stated that the survey was collected pre-COVID in the limitation section.

The second suggestion is to stress the relation of “whole – and part” and to make it clearer from the visualizations. We have added the direct relationship arrows in the proposed model to visualize and indicate the comparisons that we’ll be making between direct and indirect relationships. We have also added explanation under data analysis to further stress the comparison between full and partial mediator in this research.

Thank you for the suggestions once again.

Reviewer 3 Report

When stating the research hypothesis (2.4.), when you express a possible causal relationship between the exogenous variables and the endogenous, the expression is not correct: H1 the destination cognitive image (X) affects the overall image (Y) is how these should be expressed. It is not correct the term 'positively'; I interpret that what you are trying to express is the causal relationship between the latent variables X and Y. Another question is that the corresponding regression coefficient os this relationship is positive, meaning that a higher level of X induces higher levels of Y (and, conversely, lower levels of X imply lower of Y),

The same can be said of the expression of the rest of the hypothesis: X affects Y

Later on, when you discuss the results of the SEM model, and for the sake of statistical rigor, it is the moment to give an interpretation of the sign of the corresponding estimated coefficients (that should not be included in the hypothesis).

Regarding data, most of the variables were ordinal (for example Lickert); although it is a widely published mistake, with ordinal data, it is not possible to do arithmetic operations, such as obtaining a mean (v.gr. table 1), or correlations,..... Why don't you use statistical measures valid for ordinal data? (the median instead of the mean,...).

When you report results of the SEM model, the overall fit statistics is 543.38 with 201 degrees of freedom. And its p-value is P(chi2 > 543.38) < 0.00001, so you clearly reject the null hypothesis, which in these tests, H0 represent that the model used is adequate, while the H1 represent that the model does not fit well the data. 

Well, in this case, what you are reporting, in fact, is that you are rejecting the model proposed. That is the result is the opposite to your conclusions. So, at least, table 3 should be omitted (you can maintain RMSEA, NFI, ...., RFI, although there is no need to). Sample size can have an effect on a trend to reject the null hypothesis, but you have 311 cases and 201 degrees of freedom.

In several path diagrams, (fig. 2, 3), there seems that there are nor relations between endogenous latent variables and observable variables. Or in the final path diagram, there are only one observable variable related to each endogenous latent variables, and the presented results equate the variables OVIMAGE, INTOREV, INTOREC with Q43, Q44 and Q45, respectively; the regression coefficients are 1 and the error variance of the Q43, Q44 and Q45 are cero. So, how this should be interpreted in the model, and consider the identificability problems associated.

Author Response

Our sincerest gratitude for the comments and suggestions that are given by the third reviewer. We appreciate the detailed inputs and care given for our paper. We have adjusted the paper according to the suggestions.

All of the hypotheses have been revised to exclude the word ‘positively’. The first two rows of table 3 (Degrees of freedom and Minimum Fit Function Chi Square) have been omitted. All means have been changed to median.

As for the last concern, we are aware of the possible problems associated to using only one question to indicate one variable. However, as it has been indicated on the paper that we are testing the model proposed by Qu, et al., therefore we mimic the analysis and methodology used by Qu, et al. The research used as a reference performed the analysis using one indicator for each of the three latent variables namely overall image, intention to revisit and intention to recommend.

We hope that the revisions made based on the third reviewer’s suggestions are satisfactory. Thank you for the suggestions, we are able to create a better paper owing to the constructive comments given.